My various ignored struggles while learning
I had a moment of realization lately: the way I’ve been learning might not be the most optimal for “me”.
The Grind
I’m naturally curious. This leads mainly to two outcomes: either I learn new skills OR I spend an absurd amount of time surfing various rabbit holes. Learning is a passion of mine, but the problem is always time: how much sleep can I sacrifice to learn more or study longer?
Of course, using sleep as a bargaining chip is not a great idea: the more tired you get, the less you’re able to learn and retain anything properly.
I didn’t understand that. All I saw was that I was falling behind. The truth was that I was losing focus as the novelty wore off, or when I’d overpromise on my next actions, or got more stressed as time went by, and couldn’t learn fast enough to meet my goals.
I’m currently working on a custom RAG (Retrieval-Augmented Generation) from scratch for two main reasons: I want to understand this more deeply, and I’d love to help solve problems with that. I made some progress with the project, and I learned a lot by actively working on it instead of just theorizing.
I’m not dismissing this progress as I’m very proud of the growth. However, the problem showed its ugly head this week: did I REALLY understand what I was doing?
Sure, we all doubt ourselves at some point, but this was more specific: I was diving in head first and was making progress, but at what cost? I was reviewing the terms used and realized I didn’t really understand them fully.
The Awakening
I often dabble and figure things out: sometimes by logic or brute-force. The methods work, but they’re not always efficient.
I stumbled upon a very interesting book:
“Structure Over Chaos: How to Self-Learn like a PhD Student” by Tyler Andrew Cole.
I’d say it’s a mix of an Agile mindset and doing a research paper. I didn’t go to university or know anything about this. However, I was once again very curious.
In a nutshell, this showed me how wrong I was with my learning: I was reading a lot with the hope of finding answers instead of creating a map to find and extract the information I needed. Not reading a book from cover to cover felt like cheating. This book helped me to realign my thoughts and make sure I “ship” something instead of “trying to make things work”.
In a few days, I was able to find the various gaps in my knowledge as well as the various assumptions I had about what I thought I knew, but couldn’t really summarize it clearly.
I’m still in the early days of working this way, but since forcing myself to ship something based on a schedule is already something I know works (that good old Parkinson’s Law), the pressure I put on myself is low enough that I don’t choke under the short timeline, but I also see progress over time.
Sure, it’s not anything new since we already know about setting up a schedule to sit down and work or to have clear goals with scopes, but where it shines, in my opinion, is to bring it all together in a document to organize progress while minimizing wasted time.
So far, it gives me back hope and helped me course-correct when I was drifting towards those rabbit holes. I’m very grateful to see that more clearly now, as this is, in itself, already a great accomplishment
.


Love this reflection on structured learning. The idea of creating a knowledge map instead of grinding through everythig linearly is key, especially when working on complex stuff like RAG systems. I ran into something similar when I was trying to learn distributed systems last year, kept buidling without really understanding the core patterns and eventually had to backtrack alot.